Article, Forensic Medicine

Who is prescribing controlled medications to patients who die of prescription drug abuse?

a b s t r a c t

Background: Prescription drug-related fatalities remain a significant issue in the United States, yet there is a rel- ative lack of knowledge on the specialty-specific prescription patterns for drug-related deaths.

Methods: We designed a study that investigated Medical Examiner reports of prescription drug-related deaths that occurred in San Diego County during 2013. A Prescription drug monitoring program search was performed on each of these cases to ascertain which physician specialties had prescribed controlled substances to these pa- tients. The data were analyzed for each specialty, including pills per prescription, type of prescription, doctor shoppers (4 physicians + 4 pharmacies over 1 year), and chronic users (>=3 consecutive months of medications). Main findings: In 2013, 4.5% of all providers in San Diego County wrote a prescription for a patient who died a prescription-related death. There were a total of 713 providers who prescribed 4366 medications totaling 328928 pills. Overall, emergency physicians gave the lowest number of prescriptions per provider (1.6), whereas pain management provided the highest amount per provider (12.9). Most prescriptions went to doctor shoppers (N 50%) and chronic users (95.8%). Hydrocodone was the most frequently prescribed medication to those patients whose deaths were related to prescription drugs.

Conclusions: Emergency physicians appear to provide fewer prescriptions to those patients who die due to pre- scription drugs. Emergency physicians do, however, account for a significant proportion of total providers in this study. These results highlight the need to use Prescription Drug Monitoring Program data to closely monitor prescription patterns and to intervene when necessary.

(C) 2015

Introduction

Prescription drug-related complications and fatalities continue to plague the United States. Unfortunately, the body of literature that de- scribes this issue is lacking. Specifically, there are few studies that have evaluated prescription patterns of opioids by specialty, and there are even fewer studies that have evaluated all prescription-related deaths by specialty.

According to the available literature, in 2012, United States pre- scribers collectively wrote for 82.5 opioid pain relievers per 100 persons [1]. A study that investigated medical examiner data from 2002 to 2010 in Utah found that Family Medicine physicians accounted for the greatest proportion of Opioid prescriptions and opioid-related fatalities (24.1%/ 30.2%). Anesthesia and pain management collectively provided the most opioid prescriptions per patient (mean, 12.3), although each

* Corresponding author at: 200 West Arbor Drive, Mailcode 8676, San Diego, CA 92103. Tel.: +1 619 804 2806.

E-mail address: [email protected] (C.J. Coyne).

specialty only prescribed 1% of the total opioids given by all specialties [2,3]. Emergency medicine also provided a significant portion of opioid prescriptions [4]. However, a recent study found that emergency depart- ments (EDs) prescribed 44% fewer pills per prescription and 17% less po- tent Morphine equivalents when compared to office-based practices [5]. This study highlights the importance of investigating not only the num- ber of prescriptions but also the total number of pills per prescription.

Material and methods

This is a retrospective observational study that investigated San Diego Medical Examiner reports of prescription drug-related deaths that oc- curred in San Diego County during 2013. A prescription drug monitoring program search was performed on each of these cases to ascer- tain which physicians and physician specialties had prescribed controlled substances to these patients over the previous 12 months.

The San Diego County Medical Examiner’s Office follows Govern- ment Code Section 27491, which states that an investigation be con- ducted for all unnatural deaths that include homicides, suicides,

http://dx.doi.org/10.1016/j.ajem.2015.09.003

0735-6757/(C) 2015

accidents, deaths in custody, and certain infectious diseases. The San Diego County population includes approximately 3.2 million people, ap- proximately 1% of the US population, with approximately 20000 deaths per year, of which approximately 10000 per year are investigated by the medical examiner’s office.

The medications entered into the California PDMP database include

all Schedule II to IV medications. In 2013, this did not include tramadol (Ultram), as inclusion of this drug started in August 2014. The data are obtained from pharmacy information that is uploaded into the state sys- tem on a regular basis. All major pharmacies have the software required to comply with the regulation (SB 809), which stipulates that the data be uploaded within 1 week of each prescription. Those prescriptions ob- tained from inpatient hospital pharmacies, the Veterans Administration, Military hospitals, methadone clinics, and out-of-state pharmacies are not included.

A database was created that included all deaths in San Diego County from the period of January 1, 2013, until December 31, 2013, with a cause of death attributed to prescription drugs. This database did not in- clude suicides but did include prescription drugs that were mixed with over-the-counter medications, alcohol, or illicit drugs. A PDMP inquiry was run for the 12-month period before date of death on all patients. This report included dates of prescription, names of medication, dosage, strength, number of pills, names of pharmacies, names of providers, and provider DEA numbers. A Google search was performed to determine the Medical specialty of each provider. The study authors performed this search by entering the prescriber’s name and credentials followed by “San Diego.” The first 3 search results were evaluated to ascertain the provider specialty. The data were analyzed for each specialty to determine number of prescriptions, pills per prescription, type of prescription, doctor shoppers, and chronic users. In addition, study investigators reviewed the toxicology report for these deaths to de- termine if the medications prescribed matched what the toxicology

reports demonstrated. All data were analyzed using STATA data analysis software.

The specialties were grouped as primary care/internal medicine (IM) subspecialties (cardiology, endocrinology, family practice, General practice, gastroenterology, gynecology, infectious disease, internal med- icine, nephrology, neurology, nurse practitioners, oncology, physician assistant, physical medical and rehabilitation, and rheumatology), psychiatry, Surgery (ENT, neurosurgery, ophthalmology, orthopedics, plastic surgery, podiatry, radiology, surgery, urology, and vascular sur- gery), emergency and urgent care (ED/UC), pain (anesthesia and pain medicine), and dentistry (Table 1).

The definition of doctor shopping used for this study was the “4-4- 12” model or prescriptions from 4 different physicians, plus 4 different pharmacies over a 12-month period. This definition has been used in previous studies [6]. Chronic use was described as 3 or more consecu- tive months of a prescription medication. We calculated descriptive sta- tistics for all measures noted above [7].

Results

Demographics

The San Diego Medical Examiners Office reported 254 deaths related to prescriptions in calendar year 2013. Of those, 186 (73%) had PDMP data in the 12 months before death. The cohort of 186 patients that had PDMP data were 60% male and had a mean age of 46.4 years, with a range of 19 to 73 years.

Specialist Data

Our search identified 713 providers who prescribed 4366 medica- tions totaling 328928 pills. Among these providers, primary care/IM

Table 1

Prescription and pill numbers by specialty (in order of % total Rx)

No. of providers

No. of Rx

% Total Rx

Rx/provider

Pills/Rx

No. of Rx to doctor shopper per provider

%Rx to doctor shopper per provider

No. of Rx to chronic user per provider

%Rx by specialty to chronic user

IM

141

947

21.7

6.7

72.8

3.2

47.8%

6.5

97.2%

Psychiatry

77

795

18.2

10.5

57.9

5.8

55.7%

10.0

97.1%

Family practice

118

761

17.4

6.4

69.2

2.8

43.1%

6.1

95.0%

Physician assistant

59

274

6.3

4.6

84.9

2.2

47.5%

4.4

94.9%

Nephrology

1

1

5.1

1

30

1.0

100.0%

1.0

100.0%

ED/UC

140

217

5.0

1.6

22.9

1.0

64.1%

1.2

78.8%

Pain

13

182

4.2

14

94

6.2

44.0%

14.0

100.0%

Neurology

8

221

4.1

27.6

106.9

17.1

62.0%

27.6

100.0%

Orthopedics

22

178

4.1

8.1

169

6.1

75.8%

8.0

98.9%

Physical medicine and rehabilitation

11

155

3.6

14.1

93.9

7.5

52.9%

14.0

99.4%

Nurse practitioner

22

112

2.6

5.1

93.7

3.2

62.5%

4.9

95.5%

Gynecology

6

110

2.5

18.3

98.3

1.3

7.3%

18.2

99.1%

Dentistry

33

79

1.8

2.4

18.9

1.3

55.7%

2.0

84.8%

Anesthesia

7

75

1.7

10.7

102.9

2.4

22.7%

10.7

100.0%

Podiatry

4

54

1.2

13.5

64.2

11.0

81.5%

13.5

100.0%

Surgery

15

51

1.2

3.4

56.5

1.2

35.3%

2.9

86.3%

Oncology

4

44

1.0

11

52.6

8.0

72.7%

11.0

100.0%

Rheumatology

5

36

0.8

7.2

84.6

5.0

69.4%

7.2

100.0%

Endocrinology

1

23

0.5

23

82.2

0.0

0.00%

23.0

100.0%

Neurosurgery

3

7

0.3

2.3

100

1.7

71.4%

2.3

100.0%

General practice

3

9

0.2

3

62.4

1.3

44.4%

3.0

100.0%

Gastroenterology

1

7

0.2

7

162.9

0.0

0.0%

7.0

100.0%

Cardiology

1

4

0.1

4

39.8

4.0

100.0%

4.0

100.0%

ENT

2

3

0.1

1.5

60

1.0

66.7%

1.5

100.0%

Infectious disease

3

5

0.1

1.7

62.8

0.3

20.0%

1.0

60.0%

Plastic surgery

1

2

0.1

2

9

1.0

50.0%

1.0

50.0%

Radiology

2

2

0.1

1

36

0.5

50.0%

0.5

50.0%

Urology

5

6

0.1

1.2

41.7

1.0

100.0%

1.0

100.0%

Vascular surgery

1

2

0.1

2

45

0.8

66.7%

0.8

66.7%

Unknown

3

3

0.1

1

38.7

2.0

100.0%

2.0

100.0%

Ophthalmology

1

1

0.02

1

30

0.0

0.0%

0.0

0.0%

Total Mean

713

4366

6.1

75.3

3.1

50.7%

5.9

95.8%

Pain

3%

4%

Surgery 8%

Psychiatry 11%

Primary Care 54%

Emergency/ Urgent Care 20%

Dentistry

The 5 urologists in our study also provided 100% of their prescriptions to doctor shoppers. There were 8 neurologists in the database, who wrote a mean of 17.1 prescriptions per provider that went to doctor shoppers. This was the highest outlier, with podiatry following at 11 prescriptions per provider given to doctor shoppers.

Chronic users

Chronic use was found in 128 (68.8%) of the patients. The overwhelm- ing majority of all prescriptions were given to chronic users (95.8%). The mean number of prescriptions written to chronic users per provider was

5.9. Pain specialists gave 100% of their prescriptions to these chronic users (12.9 prescriptions per provider). The number of prescriptions to chronic users per provider was 10.0 for psychiatry, 6.8 for primary care/IM, 5.2 for surgery, 2.0 for dentistry, and 1.2 for ED/UC physicians.

Prescriptions

Figure. Percent of prescribers by specialty.

comprised the majority at 54%, followed by emergency medicine (20%) and psychiatry (11%) (Figure). The percentage of providers did not nec- essarily correlate to the percentage of prescriptions and pills. Of note, al- though most providers in the database prescribed to only 1 patient who later died due to prescription medications, there were 3 providers in the database who had each prescribed to 4 of these patients. According to the DEA registry, there were approximately 16000 DEA registrants in the San Diego Area in 2013. Therefore, more than 4.5% of all providers in San Diego gave a prescription to someone who died of a prescription-related death that year.

The mean number of prescriptions written per provider (Rx/provid- er) in 1 year was 6.1, with neurology providing the highest number at

26.3 and nephrology and radiology providing the lowest at 1. Primary care/IM physicians accounted for 62.1% of all prescriptions. This includ- ed 2.6% from nurse practitioners and 6.3% from physician assistants. Psychiatry was the next highest at 18.2%, followed by surgery 7.0%, pain specialists 5.9%, ED/UC physicians 5.0%, and dentistry 1.8% (Tables 1 and 2).

The mean number of pills per prescription (pills/Rx) was 75.3, with orthopedics providing the highest at 169 and plastic surgery with the lowest at 9. After grouping specialties, we found that dentists averaged 19, emergency physicians 23, psychiatry 58, primary care/IM 79, pain

97, and surgery 123.

Doctor shopping

There were 52 patients meeting the definition of doctor shoppers of the 186 patients with PDMP data (28%). Most of all prescriptions writ- ten (50.7%) were to doctor shoppers. The percentage of doctors within each specialty that prescribed to doctor shoppers varied. Some special- ties were high above this 50.7% mean (orthopedics and podiatry), whereas some were far below (general surgery and anesthesia) (Table 1). Of all prescriptions written by surgeons, 69.3% were given to doctor shoppers, which was the highest among all specialties. The mean number of prescriptions per provider written to doctor shoppers was 3.1. Excluding the outliers discussed in the following paragraph, psychiatrists gave the greatest number of prescriptions to doctor shop- pers per provider (5.8). Emergency and urgent care physicians gave the least amount of prescriptions per provider (1.0).

There was only 1 provider in each of the specialties of cardiology, ne- phrology, and vascular surgery that prescribed to doctor shoppers. For these 3 providers, 100% of their prescriptions went to doctor shoppers.

There were 42 medications on the PDMP reports that included opi- oids, benzodiazepines, sleep aids, stimulants, and others. The percent- age of pills and percentage of prescriptions were not always the same. Opioids were 62.8% of all pills and 53.8% of all prescriptions. Sleep aids were 3% of all pills and 8% of all prescriptions. For “other medications,” the percentage of pills was 8.2%, and for prescriptions, it was 9%. The pill and prescription overall percentages were relatively the same for benzodiazepines (24.5%) and stimulants (1.3%) (Table 3).

A total of 190 patients received opioids, which included 2350 pre- scriptions with 206700 pills, which is a mean of 1088 opioid pills/pa- tient/year. Opioids were by far the largest Medication prescription and pill category followed by benzodiazepines, other, sleep aids, and stimu- lants. Primary care prescribed most of total opioid pills (69.2%), whereas dentistry prescribed the least (0.6%). Opioids comprised 91.6% of all pre- scriptions written by surgeons, 88.6% for dentists, 80.9% for pain physi- cians, and 73.5% for ED/UC physicians.

Ninety-three patients received benzodiazepines, which included 1211 prescriptions with 80686 pills, which is a mean of 868 pills/pa- tient/year. Primary care/IM prescribed most of all benzodiazepine pills (52.0%), whereas ED/UC physicians prescribed only 0.95%. Most of the prescriptions written by psychiatrists in this database were for benzodi- azepines (72.2%), although their total pill numbers did not exceed that of primary care/IM.

Seventeen patients received sleep aids, which included 356 pre- scriptions with 9901 pills, which is a mean of 582 pills/patient/year. Zolpidem (Ambien) was the most common sleep medication pre- scribed. Primary care/IM prescribed most of total sleep aid pills (57.2%), followed by psychiatry (39.4%) and pain (2.2%). Of all prescrip- tions written by primary care/IM, 2.8% were for sleep aids, which was the highest among all specialties.

Thirty-one patients received stimulants, which included 58 prescrip- tions with 4560 pills, which averages to 147 pills/patient/year. Only 3 specialties prescribed stimulants, with psychiatry accounting for most pills (57.2%), followed by primary care/IM (41.5%) and surgery (3.4%). Of all prescriptions written by psychiatrists, 5.7% were for stimulants.

The “other” category of medications included carisoprodol as the most common medication, followed by testosterone, lyrica, phenobarbitol, dronabinol, and estrogen. One hundred twelve patients received “other” medications, which included 391 prescriptions with 27081 pills. The special- ty accounting for most “other” medication pills was primary care/IM (90.7%). Of all prescriptions written by ED/UC physicians, 5.9% were for “other” medications, followed by primary care/IM (2.8%) and surgery (0.6%).

Specific medications

Hydrocodone was the most frequently prescribed medication, with the greatest number of total pills (95 821), total patients (123), and total number of prescriptions (990) (Tables 4 and 5). Primary care/IM

Table 2

Prescription and pill numbers by specialty category (in order of % total Rx)

No. of providers No. of Rx % Total Rx

% total pills

Rx/provider

Pills/Rx

No. of rx to

%Rx to

No. of Rx to

% Rx to chronic

doctor shopper

doctor

chronic

users

per provider

shoppers

user per

provider

Primary care/IM

384

2709

62.1

65.0

7.1

79

3.3

47.1

6.8

96.8

Psychiatry

77

795

18.2

14.0

10.3

58

5.8

55.7

10.0

97.1

Surgery

56

306

7.0

11.5

5.5

123

3.8

69.3

5.2

95.4

Pain

20

257

5.9

7.6

12.9

97

4.9

37.7

12.9

100

ED/UC

140

217

5.0

1.5

1.6

22.9

1.0

64.1

1.2

78.8

Dentistry

33

79

1.8

0.5

2.4

19

1.3

55.7

2.0

84.8

Total

710

4363

50.7

95.8

Mean

6

75

3.1

5.9

prescribed the greatest percentage of hydrocodone pills (63.3%), follow- ed by surgery (27.6%), pain (4.0%), psychiatry (2.5%), ED/UC physicians (1.9%), and dentistry (0.7%). The number of total prescriptions per spe- cialty followed a different pattern, due to a difference in the mean num- ber of pills per prescriptions. Of the pills prescribed by primary care/IM, 68.3% were for hydrocodone. This far exceeds the hydrocodone percent- age prescribed by surgery (13.9%), ED/UC physicians (8.6%), dentistry (3.4%), psychiatry (3.2%), and pain (2.6%).

Oxycodone was the second most frequent medication given to those pa- tients who died in relation to a prescription medication. Primary care/IM physicians and pain specialists prescribed this medication with the greatest frequency. Alprazolam was the third most commonly prescribed medica- tion (total pills). Primary care/IM distributed the greatest percentage of al- prazolam pills (66.0%), followed by psychiatry (27.1%). Clonazepam was the third most common medication in terms of number of prescriptions. Psychiatrists prescribed this medication most frequently. Primary care/IM prescribed the greatest number of morphine, hydromorphone, methadone, carisoprodol, diazepam, and zolpidem. Psychiatry prescribed the greatest number of lorazepam pills. Emergency and urgent care physicians pre- scribed the greatest number of chlordiazepoxide pills (48.5%). Pain special- ists prescribed the greatest number of fentanyl patches (36.3%).

Discussion

Deaths related to prescription drugs have become unnecessarily common [8-11]. Opioid analgesics alone are now responsible for more

deaths than motor vehicle accidents and suicides [12]. As part of an ef- fort to curb this concerning trend, we must first identify what medica- tions are most commonly linked to prescription-related fatalities and what groups are prescribing these medications most frequently.

Primary care/IM specialties generally have large numbers of pro- viders. These large prescriber bases, however, do not adequately account for the observed number of prescription-related deaths attributed to these specialties. Emergency medicine has a similar number of providers in our study at 140, compared to internal medicine at 141, yet only 217 (5.0%) of the prescriptions were given by ED/UC physicians, whereas 947 (21.7%) were provided by IM. Sixty-nine percent of the opioids in the database were prescribed by primary care/IM, whereas only 1.76% where prescribed by emergency medicine (Table 3). This suggests that although these patients are presenting to the ED, they are still receiving most their pills elsewhere. This is despite the data showing that ED/UC physicians prescribed more to doctor shoppers (64.1%) than IM (47.8%). Internal medicine did prescribe nearly 20% more frequently to chronic users (97.2%) than ED/UC physicians (78.8%), which may partial- ly account for the 4-fold difference in total prescription numbers be- tween these 2 specialties. Previous data have shown that chronic users are at increased risk for prescription-related deaths and that this risk in- creases with the number of prescriptions dispensed [13,14].

For opioids, one would expect to see the greatest number of prescrip- tions given by primary care/IM specialties, pain specialists, and surgeons, which is what we identified in our study. One would not expect, however, that psychiatry would prescribe more opioids than ED/UC and dentistry.

Table 3

Medication categories

by specialty groups

Opioids

Benzodiazepines

Sleep aid

Stimulants

Other

Total

Total

No. of patients

190

93

17

31

112

186

No. of Rx

2350

1211

356

58

391

4366

% of total Rx

53.8

24.5

8.2

1.3

9.0

No. of pills

206700

80686

9901

4560

27081

328928

% of total pills

62.8

24.5

3.0

1.4

8.3

ED/UC

ED/UC % total pills

1.8

1.0

0.3

0

4.3

1.5

ED/UC % total Rx

6.6

3.4

0.3

0

5.7

5.0

ED/UC frequencya

73.5

15.5

0.6

0

5.9

100

PC/IM

PC % total pills

69.2

52.0

57.2

41.5

90.7

65.0

PC % total Rx

68.7

45.3

61.5

32.8

77.6

62.1

PC frequency

66.9

19.6

2.8

0.9

2.8

100

Surgery

Surgery % total pills

16.7

1.3

0.6

1.3

3.4

11.5

Surgery % total Rx

10.7

1.8

0.6

3.5

7.9

7.0

Surgery frequency

91.6

2.7

0.6

0.2

0.6

100

Dental

Dental % total pills

0.6

0.2

0.3

0

1.1

0.5

Dental % total Rx

2.9

0.7

0.6

0

0

1.8

Dental frequency

88.6

9.5

0

0

0.2

100

Psych

Psych % total pills

2.0

41.2

39.4

57.2

0.5

14

Psych % total Rx

2.9

44.9

34.8

63.8

6.0

18.2

Psych frequency

8.9

72.2

0.2

5.7

0.2

100

Pain

Pain % total pills

9.7

5.0

2.2

0

0.5

7.6

Pain % total Rx

8.2

3.8

2.3

0

2.7

5.9

Pain frequency

80.9

14.6

0.1

0

0.1

100

Abbreviation: PC, primary care,

a Percentage of medication number of pill category given by a specialty compared to all pills given by that specialty.

Table 4

Percent of pills for specific medications by specialty (in order of total pills)

Total pills No. of patients

No. of Rx

ED/UC %pills

PC/IM %pills

Surgery %pills

Dental %pills

Psych %pills

Pain %pills

Hydrocodone

95821

123

990

1.9

63.3

27.6

0.7

2.5

4.0

Oxycodone

61322

100

667

1.6

76.9

8

0.7

0.2

12.6

Alprazolam

26839

39

318

0.2

66.0

0.02

0

27.1

6.7

Clonazepam

25271

44

399

0.3

41.2

0.4

0.5

55.2

2.5

Morphine

25044

32

293

0.3

79.0

8.6

0.1

0.7

11.3

Carisoprodol

20279

30

260

1.1

75.1

9

0

10.6

4.2

Lorazepam

12239

37

215

1.6

37.0

1.0

0.03

54.5

5.9

Diazepam

11846

26

142

0.6

65.8

6.1

0.1

24.3

3.1

Hydromorphone

11397

20

128

0.4

48.9

7.0

0

1.1

42.6

Zolpiderm

8431

43

306

0.4

62.7

0.7

0.3

33.4

2.5

Methadone

5654

14

66

0

100

0

0

0

0

Chlordiazepoxide

703

17

33

51.3

40.4

2.6

0

5.7

0

Fentanyl

454

13

61

0

33.9

2.2

3.3

24.2

36.3

Perhaps this is due to the increased number of prescriptions given to chronic users by psychiatrists (97.1%) or due to a subset of psychiatrists specializing in addiction or pain management. Specifically, there were 140 ED/UC providers who prescribed 1.8% of the opioids to these patients, vs 77 psychiatrists who prescribed 2.0% of the opioids in our study.

Doctor shoppers received approximately half of the prescriptions in this study but only accounted for 20.5% of the patients. Previous studies have shown that doctor shoppers are at significantly increased risk for death from prescription medications [15]. The mean number of pre- scriptions per provider to doctor shoppers was 3.1, with neurology pro- viding 17.1 prescriptions per provider to doctor shoppers, the most in this study. An increase in an individual’s doctor-shopping behavior has been shown to precede drug-related death [15,16]. Given the large pro- portion of doctor shoppers in our study on prescription-related deaths, perhaps increased vigilance is warranted on the part of providers to identify this behavior early and to provide appropriate intervention.

The overwhelming majority of prescriptions (95.8%) in our study went to chronic users, who accounted for 68.8% of the patients. It is not surprising that a majority of patients that died due to prescription drugs over the course of this study were chronic users. With prolonged use of a potentially dangerous medication, the risk of death increases ac- cordingly [15]. One of the more interesting outliers in this study in- volved neurology, which prescribed 100% of their medications to chronic users. Although there were only 8 neurologists in this study, they collectively prescribed more pills than all of the 141 ED/UC pro- viders combined. This phenomenon was not apparent in previous stud- ies [2]. The predominance of chronic users in this investigation highlights the need for medication agreements. These patient- provider contracts (if consistently re-evaluated to avoid unwarranted refills) have been shown to significantly curb medication abuse [17]. In- creased use of PDMP data helps to identify those patients at the greatest risk of addiction and allows for expeditED referrals to addiction specialists.

One of the most important features of the PDMP is improved provid- er communication. Although we have yet to develop a universally shared medical record, the PDMP allows for prescribers across all insti- tutions and specialties to access prescription drug information and to subsequently avoid potentially fatal Drug interactions as well as dupli- cate prescriptions.

Limitations

The study had several limitations. The first limitation is that the PDMP system does not include prescriptions filled by the Veterans Af- fairs, military, hospitals, methadone clinics, out-of-state pharmacies, In- ternet, and illegal prescriptions.

In evaluating the percentage of providers that were present on the PDMP reports, a total of all DEA licenses for the area were used. The in- active DEAs and Imperial County DEAs, however, could not be separated from the San Diego DEAs. Therefore, the percentage of active San Diego County physicians who wrote a prescription to a patient who died due to a prescription-related death was likely higher than the reported 4.5%. Although DEA licenses were used to identify providers, we used a Google search to assign medical specialties without checking with the specialty medical boards. Physicians were categorized as IM, family prac- tice, psychiatry, and physical medical and rehabilitation who may have been specializing in pain management. Emergency medicine and urgent care were put together because some emergency physicians practiced in urgent care settings. There may have been other specialties that prac- ticed in an urgent care setting that were not included. Physician assis-

tants and nurse practitioners were assumed to be in primary care/IM.

Many patients did not have a toxicology report on autopsy that matched their PDMP data within 2 months of death (100/254). In addi- tion, some patients had ongoing illicit drug and alcohol use and were concurrently taking additional prescription medications. Therefore, we cannot presume causation between a certain prescription drugs and

Table 5

Percent of prescriptions for specific medications by specialty (in order of total pills)

Total pills

No. of patients

No. of Rx

ED/UC %Rx

PC/IM %Rx

Surgery %Rx

Dental %Rx

Psych %Rx

Pain %Rx

Hydrocodone

95821

123

990

8.6

68.3

13.9

3.4

3.2

2.6

Oxycodone

61322

100

667

8.0

67.0

11.5

3.2

0.3

10.0

Alprazolam

26839

39

318

0.9

60.1

0.3

0

31.8

6.9

Clonazepam

25271

44

399

0.8

34.3

0.2

0.5

62.4

1.8

Morphine

25044

32

293

1.4

78.2

6.5

0.3

0.3

13.3

Carisoprodol

20279

30

260

4.2

76.2

8.9

0

7.3

3.5

Lorazepam

12239

37

215

6.5

41.4

2.3

0.9

44.7

4.2

Diazepam

11846

26

142

2.8

57.0

8.5

2.1

25.4

4.2

Hydromorphone

11397

20

128

1.6

64.1

7.0

0

0.8

26.6

Zolpiderm

8431

43

306

0.3

67.0

0.7

0.7

28.8

2.6

Methadone

5654

14

66

0

100

0

0

0

0

Chlordiazepoxide

703

17

33

48.5

42.4

6.1

0

3.0

0

Fentanyl

454

13

61

0

52.5

1.6

1.6

18.0

26.2

death. Finally, given that this study was conducted in a specific geo- graphical region, the results may not be indicative of the specialty- specific prescribing patterns in other areas of the United States.

Conclusions

In 2013, more than 4.5% of all providers in San Diego County with a DEA license wrote a prescription for a patient with a subsequent cause of death attributed to prescription drugs. Although there is no direct correlation between prescription drug distribution and death, every medical specialty should evaluate their practices, consider what other providers are prescribing, and see where improvements can be made in their own prescribing patterns. The criterion standard is to have all chronic controlled prescriptions managed by a single provider and a sin- gle pharmacy for safety. Prescription Drug Monitoring Program review is an important tool when determining the safety of prescriptions and when identifying patients who require addiction treatment. With the insights provided by this study, we suggest that all providers should make a concerted effort to evaluate their own prescribing patterns in the context of their specialty. When prescribing new medications, pro- viders should take into account what other physicians are prescribing to avoid potentially fatal drug interactions and duplicate prescriptions. With the use of PDMP data and with increased prescriber vigilance, we can avoid prescription-related fatalities and prescription medication addition, while promoting drug rehabilitation through decreased pre- scription numbers and through increased addiction medicine referrals.

Acknowledgments

The authors thank Mike Small, Program Manager, Department of Justice, California Prescription Drug Monitoring Program.

For more information on the San Diego prescription drug abuse Medical Task Force and safe prescribing recommendations/resources, please visit SanDiegoSafePrescribing.org.

References

  1. Paulozzi LJ, Marck KA, Kochenberry JM. Variations among states in prescribing of opioid pain relievers and benzodiazepines. MMWR 2014;63(26):563-8.
  2. Porucznik CA, Johnson EM, Rolfs RT, Sauer BC. Specialty of prescribers associated with prescription opioid fatalities in Utah 2002-2010. Pain Med 2014(Issue 1):73-9.
  3. Jena AB, Goldman D, Weaver L, Karaca-Mandic P. opioid prescribing by multiple providers in medicare: retrospective observational study of insurance claims. British Medical Journal 2014.
  4. Ringwalt C, Gugelmann H, Garrettson M, Dasgupta N, Chung AE, Proescholdbell SK, et al. Differential prescribing of opioid analgesics according to physician specialty for Medicaid patient with chronic non cancer pain diagnosis. Pain Res Manag 2014(19): 179-85.
  5. Menchine M, Axeen S, Plantmason L, Seabury S. Strength and dose of opioids pre- scribed from US emergency departments compared to office practices: implications for emergency department safe-prescribing guidelines. Ann Emerg Med 2014;64:S1.
  6. Katz N, Panas L, Kim M, Audet AD, Bilansky A, Eadie J, et al. Usefulness of prescription monitoring programs for surveillance–analysis of Schedule II opioid prescription data in Massachusetts, 1996-2006. Pharmacoepidemiol Drug Saf 2010;19(2): 115-23.
  7. Chou R, Fanciullo GJ, Fine PG, Adler JA, Ballantyne JC, Davies P, et al. Clinical guide- lines for the use of chronic opioid therapy in chronic noncancer pain. J Pain 2009; 10(2):113-30.
  8. McCabe SE, Cranford JA, West BT. Trends in prescription drug abuse and depen- dence, co-occurrence with other Substance use disorders, and treatment utilization: results from two national surveys. Addict Behav 2008;33:1297-305.
  9. Paulozzi LJ, Budnitz DS, Xi Y. Increasing deaths from opioid analgesics in the United States. Pharmacoepidemiol Drug Saf 2006;15:618-27.
  10. Hall AJ, Logan JE, Toblin RL, Kaplan JA, Kraner JC, Bixler D, et al. Patterns of abuse among unintentional pharmaceutical overdose fatalities. JAMA 2008;300:2613-20.
  11. Paulozzi LJ, Jones CM, Mack KA, Rudd RA, et al. Vital signs: overdoses of prescription opioid pain relievers–United States, 1999-2008. MMWR 2011;60:1-6.
  12. Manchikanti L, Helm S, Fellows B, Janata JW, Pampati V, Grider JS, et al. Opioid epi- demic in the United States. Pain Physician 2012;15(3 Suppl):ES9-S38.
  13. Ekholm O, Kurita GP, Hojsted J, Juel K, Sjogren P. Chronic pain, opioid prescriptions, and mortality in Denmark: a population-based cohort study. Pain 2014;155(12):2486-90.
  14. McDonald DC, Carlson EK. Estimating the prevalence of opioid diversion by “doctor shoppers” in the US. PLoS One 2013. http://dx.doi.org/10.1371/journal.pone 0069241.
  15. Peirce GL, Smith MJ, Abate MA, Halverson J. Doctor and pharmacy shopping for con- trolled substances. Med Care 2012;50(6):494-500.
  16. Martyres RF, Clode D, Burns JM. Seeking drugs or seeking help? Escalating “doctor shopping” by young Heroin users before fatal overdose. Med J Aust 2004;180:211-4.
  17. Bujold E, Huff J, Staton EW, Pace WD. Improving use of narcotics for nonmalignant chronic pain: a lesson from Community Care of North Carolina. J Opioid Manag 2012;8(6):363-7.